On Thu, Sep 28, 2023, Maxim Levitsky wrote: > У чт, 2023-09-28 у 12:29 -0400, Paolo Bonzini пише: > > The "bool shared" argument is more or less unnecessary in the > > for_each_*_tdp_mmu_root_yield_safe() macros. Many users check for > > the lock before calling it; all of them either call small functions > > that do the check, or end up calling tdp_mmu_set_spte_atomic() and > > tdp_mmu_iter_set_spte(). Add a few assertions to make up for the > > lost check in for_each_*_tdp_mmu_root_yield_safe(), but even this > > is probably overkill and mostly for documentation reasons. > > Why not to leave the 'kvm_lockdep_assert_mmu_lock_held' but drop the shared > argument from it? and then use lockdep_assert_held. If I am not mistaken, > lockdep_assert_held should assert if the lock is held for read or write. +1, I don't see any downside to asserting that mmu_lock is held when iterating. It'll be a redundant assertion 99% of the time, but it's not like performance matters all that much when running with lockdep enabled. And I find lockdep assertions to be wonderful documentation.
